Thao Phan, Monash University; Fabian Offert, University of California Santa Barbara;
What is synthetic media? While born-digital objects have existed since the 20th century, the current AI revolution has put a new spin on the question of digital remediation, in both scientific and cultural knowledge production.
Large visual models are able to emulate established media like photography without simulating any particular optical apparatus. Large language models are able to extrapolate text based on minor prompts, an approach that literalises the cybernetic maxim of communication as information processing. In what is either the dream or nightmare of media archaeology, established media become infinitely remixable, as just one 'style' among others.
Synthetic media, then, represents a turn from an aesthetics and politics governed by semiotic, ideological and other forms of symbolic meaning, to an aesthetics and politics driven by statistical correlations and probabilistic determinations. Here, the 'stuff' of culture, such as images and words, are not produced through systems of human meaning and intention but through opaque computational processes that ontologically flatten the ineffable qualities of culture into operational data points.
This panel invites STS scholars interested in these new ways of seeing, listening, reading, producing, consuming and other ways of being with media. It includes topics like large language and visual models, but also extends to practices such as the use of artificially manufactured datasets (synthetic data) and artificial environments (training simulations). We are especially keen to receive submissions that engage with questions of aesthetics and politics as articulated through feminist, critical race, and other frameworks grounded in material analyses of power.